What is the Maximum Capture Bitrate i should use ??

[Nelson37]

For MMC, playback was authored capture burned to disk and played through DVD player. TNTHD was viewed on same TV as originally broadcast, I did not compare with a capture. of TNTHD

DVD player has component connection, satellite has HDMI. 42" LCD.

Simple visual check. If I can't see it, I don't care about it.

Errors generally visible as a pause or jerk forward. Any errors near possible edited adult content ignored. Most folks who have not been watching for these types of issues for years probably would not even notice it. Durations of fractions of a second.

These types of errors are to me quite common, roughly one per hour as a general rule. Sometimes these are for special effects purposes and are in the original film. An example is Star Wars :ANH when they are in the prison block rescuing the Princess. Watch closely when the doors rapidly open and close, there is a noticeable freeze in the film. I do not consider such an effect an "error", either way it is beyond my control and nothing I can do about it.
This is about equally noticeable to the framerate issues I mentioned. IMO very, very minor.

The TNT SD channel uses some of the most godawful screwed up Telecine patterns I have ever seen, never had an effective software or hardware combination to handle it. Tried to cap The Mummy at least four times. I did save a clip of that babe in the net suit walking down the hall. Damn!

Their HD channel does not seem as bad, though I have made no detailed analysis and no attempt at IVTC, due to the commercial interruptions.

Your question brings up a question I forgot to ask.

Has anyone else besides me actually tried the ATI Inverse Telecine on appropriate sources as discussed here? More than once?

The only "Cadence Analyzer" I know of is the one built into TMPGENC, which does actually analyze the video and give some output on the pattern. I do not recall outstanding success using this program for IVTC purposes, it may just have been exceptionally slow compared to Avisynth filters.

[c627627]

I have a situation that I can replicate that made me remember why I said to select 'Progressive Source' and not 'Encode Interlaced' if capturing a high definition signal.

I have an HD show that I captured into an AVI file.

The opening sequence is jittery if 'Interlaced' is selected for conversion to MPEG-2.
It is stable if 'Progressive Only' is selected.


What is the explanation for this high definition signal capture only being stable stable if 'Progressive Only' and not 'Interlaced' is selected?

EDIT:

For the next four pages I struggle to finally understand this:

1. I usually capture with MMC directly to mpg with Encode Interlaced then convert to DVD.
2. Capturing this show to mpg with Encode Interlaced resulted in unstable capture.
3. Capturing this show to mpg with Progressive Source resulted in a seemingly stable capture (but at the expense of everything we talked about.) I was wrong to think that just because it seemed more stable than Encode Interlaced, that's why it should be captured like that.

I thought I only had Encode Interlaced, Deinterlace and Progressive options, period. I was unaware of other problems that can exist.

4. I then realized I needed to edit this show more so I captured it to avi and erased it from my DVR.

THEN:

5. AVI which has no flags was converted to MPG with the same results as when I captured the show to mpg. Encode Interlaced was unstable no matter what.


THEN FINALLY eureka:

6. The source AVI inside the program it is loaded into such as Sony Vegas has properties as a source that can be changed before encoding. That solves the problem.


THEN AGAIN eureka:
If I wanted to capture to mpg, I could have:
1. Used DGMPGDec to demux the mpg
2. Used ReStream to UNCHECK 'top field first'

That also fixes the problem of the unstable Encoded Interlaced mpg.

[jagabo]

If the picture shook 30 times a second when there was motion you probably had the wrong field order.

[c627627]

I am now retesting all 3 options.
Interlaced Top Field First
Interlaced Bottom Field First
Progressive Only

[c627627]

Yup: Progressive Only is stable.

Both Interlaced Top Field First and Interlaced Bottom Field First are jittery like when you select a wrong field order.


This is an avi file captured with an S-Video cable from an HD cable box outputting an HD show.
Why is encoding as Progressive Only stable?


I wish I had that show to test capturing it directly as Encode Interlaced mpg to see what would happen.

[edDV]

I am now retesting all 3 options.
Interlaced Top Field First
Interlaced Bottom Field First
Progressive Only

The only difference between 480i and 1080i as far as telecine is concerned is the later was likely remastered in the last 7 years on better equipment. Also, more care is used with HD mastering because these tapes need to feed MPeg encoders and hardware downscalers for simultaneous 1080i, 720p, 480p and 480i feeds. Bad telecine cadence will upset these encoders and downscalers.

Almost everything you currently see was edited at 29.97 (separate online edits are done at 25fps for PAL releases). It has only been in the last year that Avid and others have released software for 23.976 on-line editing from 1080i/29.97 film transfer tapes. The idea with that is one 23.976 edit can be used to derive masters for both NTSC and PAL markets. Progressive editing can be done without worries about 5 frame cadence.
http://www.zerocut.com/tech/p24.html
http://phx.corporate-ir.net/phoenix.zhtml?c=82844&p=irol-newsArticle_Print&ID=696690&highlight=
http://www.avid.com/solutions/postproduction/advanced_workflows/
http://www.avid.com/onlinesupport/supportcontent.asp?productID=114&contentID=9202&typeID=

Otherwise 480i and 1080i have the same issues. Show openings are likely to be full of special effects.

[c627627]

Weird. Loaded the .avi capture into Canopus ProCoder 1.5

Sure thing, Non-Interlaced setting was the only one that came out in the converted .mpg as stable.

My original results were from Sony Vegas and now Canopus ProCoder has the same thing happening.


Hm... what would have happened if I captured it not as avi but as an .mpg directly, what would have happened if I set Encode Interlaced vs. Progressive Source at the time of capture?

[c627627]

I posted before reading edDV's reply.

I don't understand your post.


We said to always Encode Interlaced, didn't we? Why is the file only coming out as stable if encoded to .mpg with the Progressive Only setting?


Let me rephrase the questions:

1. This file was captured as an avi file. If I captured it as .mpg Encode Interlaced would it have also been jittery?

2. If instead of Encode Interlaced, "Progressive Source" was selected when capturing to .mpg, would it then not be jittery?

[edDV]

I posted before reading edDV's reply.

I don't understand your post.


We said to always Encode Interlaced, didn't we? Why is the file only coming out as stable if encoded to .mpg with the Progressive Only setting?


Let me rephrase the questions:

1. This file was captured as an avi file. If I captured it as .mpg Encode Interlaced would it have also been jittery?

2. If instead of Encode Interlaced, "Progressive Source" was selected when capturing to .mpg, would it then not be jittery?

You are viewing this on a TV from an interlace DVD?

I don't have MMC installed currently. I'm rebuilding the ATI-AIW machine. Otherwise I would try to duplicate what you are doing.

[c627627]

I know what wrong field order looks like because before finding out that my card was "Top Field First" I made a mistake of selecting "Bottom Field First" when making DVDs and those would be jittery.

I noticed that when I played those wrongly made DVDs using PowerDVD program on the computer, PowerDVD would replicate all the jitters so I noticed I could use PowerDVD to check how DVDs would look on my DVD player in terms of jitters.

Now I can make a DVD out of all these files if you want but here are the facts:

avi file encoded to DVD is jittery when played on PowerDVD unless it is encoded as Progressive Only


I also seem to remember that when I captured this show directly to mpg:

Selecting 'Encode Interlaced' would capture jittery .mpg files.

Selecting 'Progressive Only' without checking Inverse 3:2 Pulldown would capture stable .mpg files and this would only happen if I captured from an HD channel through an S-Video cable.


Those files would be jittery when viewed from my DVD player on TV.

[c627627]

The jittery parts have massive camera movements all over the place.

[edDV]

Let me rephrase the questions:

1. This file was captured as an avi file. If I captured it as .mpg Encode Interlaced would it have also been jittery?

It should make no difference for uncompressed interlace avi vs. interlace MPeg2 so far as "jitter". The fact that you asked in this contex makes me suspect you are viewing on a progressive computer monitor.

If so, you will see some jitter as frames 4 and 5 take twice the time to play as a progressive frame 4. If you look very closely you would see frames 1, 2 and 3 appear to play faster than frames 4 and 5 in the sequence. This property is part of the telecine territory.

A "cinema" processing progressive DVD player will do the IVTC back to four frames at 23.976 rate and then repeat the four frames as AABBBCCDDD at 59.94 fps. This motion will appear smoother than normal 29.97 telecine and is one of the benefits of having a progressive TV.

2. If instead of Encode Interlaced, "Progressive Source" was selected when capturing to .mpg, would it then not be jittery?

It would not be as jittery on the computer monitor if a full successful IVTC was performed. That would give you a ABCD 23.976 sequence that the display card would refesh at 72fps or whatever. Maximum smoothness would be at 71.928 or 95.904 fps screen refresh.

If you play normal telecine 29.97 to a computer monitor without IVTC, you will see frames 1, 2 and 3 clear and 4 and 5 blurred since their fields derive from different frames. At full 1x speed it takes a sharp eye to see these differences. On a normal TV, fields are displayed sequentially so you don't get this frame 4 and 5 blur to the same degree. The eye smooths it out.

[edDV]

You made a mistake thinking PowerDVD is a good simulation for a progressive DVD player feeding 480p to a progressive TV. This adds another layer of complication.

The software players do reasonably well playing a properly authored commercial DVD with good cadence flags but are not good as a class dealing with TV captures that don't have the flags.

Read the reviews here:
http://www.hometheaterhifi.com/cgi-bin/shootout.cgi?function=search&articles=106

The discussion so far has been focused on how to make the picture best on a standard TV or a progressive HDTV. Computer display is more complicated and variable depending on the player and hardware.

[c627627]

Well allright, I'll make actual real DVDs, then play them on my DVD player to shift the discussion to how Encode Interlaced vs. Encode Progressive Source looks on a DVD player and TV.

I am saying that if Encode Interlaced is selected, it's as if a DVD is made with the wrong field order.

Will post in 10 minutes with DVD player results.

[edDV]

Well allright, I'll make actual real DVDs, then play them on my DVD player to shift the discussion to how Encode Interlaced vs. Encode Progressive Source looks on a DVD player and TV.

I am saying that if Encode Interlaced is selected, it's as if a DVD is made with the wrong field order.

Will post in 10 minutes with DVD player results.

I'll try to get MMC running again in the next few days. I see no problem capturing 480i interlace off my Canopus ADVC from a Motorola HD cable box S-Video output and authoring an interlace DVD from Mainconcept based encoders. I haven't looked closely at ATI captures.

All the DVD player tests at hometheaterhifi.com are from interlace input with and without telecine.

[c627627]

avi file captured from an HD channel encoded to .mpg with Interlaced Top Field First option from which a DVD disc was made:

Jittery when viewed on the DVD player and TV as if a wrong field order was selected.


So there's your proof since I made a DVD disc from the .mpg file that was encoded with Progressive Only setting and it was stable.



Now:

Our conclusion to always use Encode Interlaced setting is therefore incorrect, agreed?


If this file were to be captured not with Encode Interlaced setting but with Progressive Source without checking Inverse 3:2 pulldown the captured file would make a stable DVD.

Therefore our other conclusion that Progressive Source should only be used in combination with checking Inverse 3:2 Puldown is also incorrect, agreed?

[c627627]

Note that I did not see jiittering when I encoded another high definition show with both Encode Interlaced as well as Progressive Only settings and made DVDs out of both.

The particular capture in question had MASSIVE camera shakes and movements that only make a stable DVD disc if encoded as Progressive Only, not if encoded as Encode Interlaced.

[edDV]

I've found some good video to use for tests. I've got the Northern Exposure DVD's (encoded progressive) and have some of the same episodes captured off Universal HD as 720x480i with telecine.

The true double field mark of telecine source.

[edDV]

Does this sample have telecine or not? If it encodes as if field order is wrong, then field order is probably set wrong in the MMC capture and the sample probably is not film. How are you authoring the DVD? Which software? Are you sure it isn't re-encoding?

[c627627]

Right but you only need one pink zebra to prove there are pink zebras in the world, right?

I had the thing on my DVR and the one & only setting that was changed to make a stable capture was 'Progressive Source' instead of 'Encode Interlaced'.


If, I say IF this is true, then it can be said that there are cases when 'Progressive Source' must be selected to get a stable capture,

but those cases would be source dependent and not whether a channel is SD or HD, correct?


More importantly how would I test for this?

[c627627]

I just read your second post, I'll reply.

[c627627]

I will take consecutive frame captures and let you decide if it has telecine.

It is not re-encoding because it take less time to encode than the sample is long. I tested Sony Vegas and ProCoder 1.5 with same results.

There are no field order settings in MMC when .avi files are captured.


I do seem to remember capturing the same show to MPEG-2 DVD setting in MMC with results being Encode Interlaced jittery and Progressive Source stable.

[edDV]

I will take consecutive frame captures and let you decide if it has telecine.

It is not re-encoding because it take less time to encode than the sample is long. I tested Sony Vegas and ProCoder 1.5 with same results.

There are no field order settings in MMC when .avi files are captured.


I do seem to remember capturing the same show to MPEG-2 DVD setting in MMC with results being Encode Interlaced jittery and Progressive Source stable.

Well I assure you that all the HD channels are playing smooth as 1080i and 480i so if it is playing jittery something is wrong with MMC or the settings.

Progressive on a non-film source will do a field blend deinterlace. Has nothing to do with smooth playback. If you don't deinterlace, it will play smoothly as well so long as field order isn't reversed.

[c627627]

12 consecutive frames:

test.jpg

[c627627]

1. What is the theoretical explanation for Progressive Source captures being stable vs. Encode Interlaced captures being jittery?

2. Even if something is wrong, this is source dependent, do you think that it has to do with SD vs. HD source or that SD vs. HD has nothing to do with it.

[edDV]

That looks like a non-film blend deinterlace. What was the program? Looks like a news promo.

1. What is the theoretical explanation for Progressive Source captures being stable vs. Encode Interlaced captures being jittery?

2. Even if something is wrong, this is source dependent, do you think that it has to do with SD vs. HD source or that SD vs. HD has nothing to do with it.

Progressive vs interlace on normal video should not affect smoothness other than a field reversal. "Progressive" will just blur the fields together.

Has nothing to do with 480i SD vs 1080i HD.

[c627627]

It was a PBS documentary type of thing, it was called Status Anxiety.

Now we're getting somewhere!

• If the source is non-film blend deinterlace, select 'Progressive Source' instead of 'Encode Interlaced'.
(Uncheck: Inverse 3:2 Pulldown)


OK, how is non-film blend deinterlace defined and recognized please.

No links and short answer please

[edDV]

That is a tough example. It doesn't show up in PBS A-Z. Too obscure. Hard to find out how it was shot but looks like a TV production not film.

If you have a source that is straight 480i video, interlace encoding should look the same field by field with the original. The only differences will result from MPeg compression. I frames should look the same.

A "Progressive" setting on 480i source will force a deineterlace. Field one will be blended with field two removing any motion difference between field one and two.

[edDV]

http://www.kqed.org/tv/indieproducers/pdfs/statusanxiety.pdf

A really bad example because it was shot in UK in PAL and transcoded to NTSC

Take a program off the PBS HD channel. Try to find one shot in film like a nature show. You are more likely to get a film source with telecine on network drama or movies or drama on HBO, TNT, UHD

[c627627]

Yeah, I looked for it too, it's not airing in the next two weeks which is how far HD schedules go at PBS. But we already established that it is a rare example so now all we need to do is account for it.

Hm... There is a De-Interlace capture setting but the confusion here stems again from not doing the simple things first:
I'm going to number the questions for you:

1. Is the name for my unusual example "non-film blend deinterlace," if not, what is the name for it and how is it defined (remember telecine is three clean frames followed by interlacing artifacts on the 4th and 5th frame) I need a one sentence description like that.

2. Once the rare source is recognized there are two and only two alternatives capture settings for it:
De-Interlace and Progressive Source so the simple thing to do is to say which to select, I'm guessing Progressive Source.

3. Leave Inverse 3:2 Pulldown unchecked, correct?